Oxygen permeability and ionic conductivity of perovskite-related La0.3Sr0.7Fe(Ga)O3-δ

Article abstract of DOI:10.1149/1.1459715

The oxygen ionic conductivity of La_0.3Sr_0.7Fe_1-xGa_x O_3-δ (x = 0.2-0.4), determined by oxygen permeation, faradaic efficiency, and total conductivity measurements at 1023-1223 K, is essentially independent of oxygen chemical potential and structural changes in the wide oxygen partial pressure range from 10^-14 to 0.21 kPa. At oxygen pressures close to atmospheric air, the ion transference numbers of perovskite-like La_0.3Sr_0.7Fe(Ga)O_3-δ phases vary from 4 × 10^-4 to 4 × 10^-2, increasing with gallium content and temperature. Although there is a great difference between ionic and p-type electronic conductivities in oxidizing atmospheres, the electron-hole conduction was demonstrated to affect oxygen permeation and ambipolar conductivity. In oxidizing conditions, the oxygen permeability of La_0.3Sr_0.7Fe_1-xGa_x O_3-δ (x = 0-0.4) membranes increase with increasing x. The bulk ionic transport process and surface exchange kinetics both influence the permeation through La_0.3Sr_0.7FeO_3-δ ceramics, whereas the oxygen fluxes through Ga-containing materials are predominantly determined by the bulk ambipolar conduction. Thermal expansion coefficients of La_0.3Sr_0.7Fe_1-xGa_xO _3-δ (x = 0-0.4) in air vary in the range (11.7-14.9) × 10⁶ K^-1 at 300-800 K and increase up to (19.5-26.4) × 10⁶ K^-1 at 800-1170 K. Substitution of iron with gallium was found to suppress both thermal expansion and chemically induced expansion of La_0.3Sr_0.7Fe(Ga)O_3-δ materials, originating from oxygen nonstoichiometry variations under changing the oxygen partial pressure.

Full text of DOI:10.1149/1.1459715

Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
E125
0013-4651/2002/149͑4͒/E125/11/$7.00 © The Electrochemical Society, Inc.
Oxygen Permeability and Ionic Conductivity
of Perovskite-Related La0.3Sr0.7Fe„Ga…O3À␦
b
a,c
V. V. Kharton,a,b,z A. A. Yaremchenko,b, A. P. Viskup, M. V. Patrakeev,
*
I. A. Leonidov,c V. L. Kozhevnikov,c F. M. Figueiredo,a,d A. L. Shaulo,b
b,**
a,**
E. N. Naumovich,
and F. M. B. Marques
aDepartment of Ceramics and Glass Engineering, Research Unit on Ceramic Materials,
University of Aveiro, 3810-193 Aveiro, Portugal
bInstitute of Physicochemical Problems, Belarus State University, 220080 Minsk, Belarus
cInstitute of Solid State Chemistry, Ural Division of The Russian Academy of Sciences,
Ekaterinburg 620219, Russia
dDepartment of Science and Technology, Universidade Aberta, 1269-001 Lisbon, Portugal
The oxygen ionic conductivity of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4), determined by the oxygen permeation, faradaic effi-
ciency, and total conductivity measurements at 1023-1223 K, is essentially independent of oxygen chemical potential and struc-
tural changes in the wide oxygen partial pressure range from 10Ϫ14 to 0.21 kPa. At oxygen pressures close to atmospheric air, the
ion transference numbers of perovskite-like La0.3Sr0.7Fe͑Ga͒O3Ϫ␦ phases vary from 4 ϫ 10Ϫ4 to 4 ϫ 10Ϫ2, increasing with
gallium content and temperature. Although there is a great difference between ionic and p-type electronic conductivities in
oxidizing atmospheres, the electron-hole conduction was demonstrated to affect oxygen permeation and ambipolar conductivity. In
oxidizing conditions, the oxygen permeability of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0-0.4) membranes increase with increasing x. The
bulk ionic transport process and surface exchange kinetics both influence the permeation through La0.3Sr0.7FeO3Ϫ␦ ceramics,
whereas the oxygen fluxes through Ga-containing materials are predominantly determined by the bulk ambipolar conduction.
Thermal expansion coefficients of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0-0.4) in air vary in the range (11.7-14.9) ϫ 106 KϪ1 at 300-
800 K and increase up to (19.5-26.4) ϫ 106 KϪ1 at 800-1170 K. Substitution of iron with gallium was found to suppress both
thermal expansion and chemically induced expansion of La0.3Sr0.7Fe͑Ga͒O3Ϫ␦ materials, originating from oxygen nonstoichiom-
etry variations under changing the oxygen partial pressure.
© 2002 The Electrochemical Society. ͓DOI: 10.1149/1.1459715͔ All rights reserved.
Manuscript submitted February 12, 2001; revised manuscript received October 18, 2001. Available electronically March 12, 2002.
Ceramic membranes with mixed oxygen ionic and electronic
conductivity are of significant interest for high temperature electro-
chemical applications, including separation of high-purity oxygen
from air, partial oxidation of methane, and other light hydrocarbons,
waste reduction and recovery, and sensors.1-7 Along with the infinite
theoretical selectivity with respect to oxygen, the mixed-conductive
membranes enable integration of oxygen separation, steam reform-
ing, and partial oxidation into a single step for the natural gas con-
version. This technology may result in considerable economical ben-
efits, since the cost of an oxygen plant constitutes the most
significant cost associated with synthesis-gas generation by oxida-
tion of methane. Moreover, a favorable feature for the hydrocarbon
oxidation process is that the oxygen flux may alter the relative
amount of different active oxygen species on the membrane surface,
thereby providing higher selectivity for partial oxidation reactions.1,2
However, the membrane materials known at the present time
exhibit various specific disadvantages that limit, essentially, their
applicability. In particular, perovskite-like solid solutions
(A,Ln)(Co,Fe)O3Ϫ␦ ͑where A and Ln are the alkaline-earth and
rare-earth elements, respectively͒ possess the highest oxygen perme-
ability in oxidizing conditions, but are thermodynamically and/or
dimensionally unstable under large oxygen chemical potential gra-
dients typical for the membrane reactor operation.6-10 In addition, at
temperatures below 1000-1070 K the most-permeable oxide mixed
conductors are characterized by ordering in the oxygen sublattice
and a high reactivity with gaseous species such as CO2 , that leads to
ported to be a promising approach to improve structural stability of
these phases under large oxygen chemical potential gradients and to
preserve a satisfactory level of the ion conductivity at the same
time.7,10,23 For instance, considerable oxygen permeation fluxes
and
a sufficient stability were found for the compositions
Sr1ϪxLaxFe1ϪyGayO3Ϫ␦ ͑x ϭ 0.1-0.4; y ϭ 0.2-0.5͒, in which the
crystal lattice was identified as brownmillerite-type.7 Notice, how-
ever, that the electroneutrality condition in the latter compounds
requires a significantly larger oxygen content than that characteristic
of the brownmillerite, Ca2AlFeO5 , especially in oxidizing condi-
tions when a considerable part of iron ions exists in tetravalent
state.20 Our previous study24 of the system La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦
(0 р x р 0.5) showed that the solid solution formation range in
both oxidizing and reducing environments corresponds to the x val-
ues from 0 to ϳ0.3. At oxygen partial pressures above 10Ϫ4 atm,
these compounds possess perovskite-type structure, whereas de-
creasing oxygen chemical potential leads to various types of
oxygen-vacancy ordering. Therewith, the crystal structure of the re-
Ͼ
duced form of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x
0) cannot be inter-
preted as a brownmillerite one.24 As ordering in the oxygen sublat-
tice may significantly affect oxygen ionic conductivity of the title
materials, particular emphasis in this work was given to comparison
of the ionic transport properties in oxidizing and reducing
atmospheres.
Experimental
Dense
ceramic
samples
of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦
a degradation in membrane performance with time.10-14
(x ϭ 0-0.40) were prepared by a standard solid-state synthesis
route. Before weighting of the stoichiometric amounts of high-purity
SrCO3 , La2O3 , FeC2O4•2H2O, and Ga2O3 , the oxides of lantha-
num and gallium were annealed in air at 1270 K for 3-4 h. The
starting mixtures were dissolved in an aqueous solution of nitric
acid, and then dried and calcined at moderate temperature to form
fine, homogeneous powders. The solid-state reactions were con-
ducted in air at 1470 to 1640 K for 20-40 h with multiple interme-
diate grinding steps. Dense ceramic samples were pressed at 250-
400 MPa and then sintered in air for 3-5 h. The compositions with
The present work, continuing our research on mixed-conductive
membrane materials,6,14-22 is focused on ionic transport properties of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ ͑LSFG͒ ceramics. Incorporation of gallium
into the crystal lattice of perovskite-related (Sr,La)FeO3Ϫ␦ was re-
* Electrochemical Society Student Member.
** Electrochemical Society Active Member.
z Email: kharton@cv.ua.pt
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E126
Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
where to is the oxygen ion transference number, and ␴, ␴o , and ␴
e
Table I. Abbreviations and properties of La0.3Sr0.7Fe1ÀxGaxO3À␦
ceramics.
represent the total, oxygen ionic, and electronic conductivities, re-
spectively. Here, ␴ is the sum of the n-type and p-type contribu-
e
Relative
density
Unit cell
parameters
a
Sintering
temperature
͑Ϯ10 K͒
͑K͒
tions (␴ ϭ ␴n ϩ ␴p). When oxygen exchange influences the over-
e
all ion transfer, the value of J(O2) should increase with membrane
thickness at a given oxygen chemical potential gradient owing to a
decreasing role of the surface exchange.
␳
exp /␳
theor
x
Abbreviation
͑nm͒
͑%͒
In this paper the reported values of the activation energy (Ea) for
either electronic or ionic conductivity were calculated using the
standard Arrhenius model
0
0.20
0.40
LSFa
LSFG20
LSFG40
0.38731
0.38828
0.39024
1560
1650
1690
93.4
92.6
92.9
a Composition is La0.3Sr0.7FeO3Ϫ␦ (x ϭ 0).
A0
T
Ea
␴ ϭ
exp Ϫ
͓3͔
ͩ
ͪ
RT
x ϭ 0, 0.20, and 0.40 were selected for detailed studies of the ion
transport properties; their abbreviations and sintering temperatures
of gastight ceramics are listed in Table I. The density of the ceramic
materials (␳exp) determined by the standard picnometric technique
was higher than 92% of their theoretical density (␳theor), calculated
from the results of X-ray diffraction ͑XRD͒ studies ͑Table I͒. Prior
to further characterization, the sintered samples were annealed in air
at 1170 K for 4 to 5 h and slowly furnace cooled in order to obtain
oxygen nonstoichiometry as close as possible to equilibrium at room
temperature.
Characterization of the materials was carried out by XRD, scan-
ning electron microscopy and energy dispersive spectroscopy ͑SEM/
EDS͒, emission spectroscopic analysis, thermal gravimetric and
differential thermal analysis ͑TGA/DTA͒, infrared absorption spec-
troscopy ͑IRAS͒, coulometric titration, and dilatometry. The experi-
mental procedures and equipment used for physicochemical charac-
terization, including coulometric titration,25 were reported elsewhere
͑Ref. 15-20, 24-28, and references therein͒. Studies of the transport
properties included four-probe dc measurements of the total conduc-
tivity ͑TC͒, determination of steady oxygen permeation fluxes ͑OP͒,
and faradaic efficiency ͑FE͒. A detailed description of the experi-
mental technique and equipment for measurements of the total con-
ductivity as a function of the oxygen partial pressure is found in Ref.
26. The details on the oxygen permeability and faradaic efficiency
measurements were also published earlier.6,15,19,28
where A0 is the pre-exponential factor.
Results and Discussion
Characterization of the ceramic materials.—XRD studies of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ powders and ceramics, equilibrated with at-
mospheric air, showed formation of a single perovskite-type phase
within the composition range 0 р x р 0.3. Further substitution of
iron with gallium leads to the segregation of the secondary phase
SrLaGa3O7 .30 The intensity of the strongest reflections of the
SrLaGa3O7 phase in the XRD patterns of La0.3Sr0.7Fe0.6Ga0.4O3Ϫ␦
did not exceed 4-5% of the most intensive reflection of the perov-
skite. Crystal structure of the perovskite-type phases was prelimi-
nary identified as cubic. It should be noted, however, that a slight
broadening of peaks at high 2␪ angles in the XRD patterns of all
studied compositions may indicate a crystal symmetry lower than
primitive cubic. For instance, the perovskite structure of the oxi-
dized form of the Sr2LaFe3O9Ϫ␦ compound is rhombohedrally
distorted.31 A more detailed structural refinement is now in progress.
The perovskite unit cell parameters of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ in-
crease with gallium concentration ͑Table I͒. Similar behavior
was reported for Fe-doped La(Sr)Ga(Mg)O3Ϫ␦ perovskite
solid solutions32 and (Sr,La)(Fe,Ga)O3Ϫ␦ phases claimed as
brownmillerites;7 possible reasons for such an increase were ana-
lyzed elsewhere.24
The permeation and faradaic efficiency tests, as well as isother-
mal measurements of the total conductivity as a function of the
oxygen partial pressure were carried out in the temperature range
1023-1223 K. For all data on oxygen permeability presented in this
paper, the membrane feed-side oxygen partial pressure (p2) was
maintained at 21 kPa ͑atmospheric air͒. All faradaic efficiency re-
sults were obtained in air under zero oxygen chemical potential
gradient.
Figure 1 shows SEM micrographs reflecting microstructures that
are typical for the La0.3Sr0.7Fe(Ga)O3Ϫ␦ ceramics. The average
grain size increases with increasing x from 7-10 ␮m, characteristic
of La0.3Sr0.7FeO3Ϫ␦
, to 18-30 ␮m observed in the case of
͑LSFG40͒. SEM studies also indicated traces of a liquid phase for-
mation at the grain boundaries of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ ceramics
at x у 0.2. Probably, both these features result from the increase in
the sintering temperature necessary to obtain gastight ceramics of
Ga-doped lanthanum-strontium ferrites ͑Table I͒.
In this article, the oxygen permeation processes are discussed
using the quantities permeation flux density j ͑mol sϪ1 cmϪ2͒ and
specific oxygen permeability J(O2) ͑mol sϪ1 cmϪ1͒ interrelated as29
The total conductivity of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ , predomi-
nantly p-type electronic in air,24 decreases with increasing gallium
concentration in the B sublattice of the ABO3 perovskite structure
͑Fig. 2͒. Evidently, the respectively decreasing amount of the Fe-
O-Fe bonds, over which the electron conduction in the ferrite lattice
occurs, is a main reason for this behavior. Another consequence of
the Ga doping is the increase in the activation energy for the p-type
conductivity at low temperatures ͑Table II͒, probably associated
with the decrease in covalency of the B-O-B bonds. When iron
cations are replaced with insulating gallium ions, the absolute con-
centration of Fe4ϩ decreases with decreasing total iron concentra-
tion. As a result, the average oxidation state of the B cations and,
hence, covalency of the B-O-B bonds, decrease.33 This seems to
favor narrowing of the conduction band and progressive localization
of electron states.33 The conductivity of La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦ at
temperatures above 600 K is very close to the literature data10 on
La0.5Sr0.5Fe0.8Ga0.2O3Ϫ␦ solid solution, while at lower temperatures,
the total conductivity of the latter compound becomes smaller ͑Fig.
2͒. As the Fe:Ga concentration ratios in these phases are equal, the
p-type electronic conduction is governed by the La:Sr ratio. Increas-
Ϫ1
p2
J O ͒ ϭ jd ln
͓1͔
͑
ͫ
2
ͬ
p1
where d is the membrane thickness, and p1 is the oxygen partial
pressure at the membrane permeate side. The quantity J(O2) is suit-
able to identify a limiting effect of surface exchange rate on the
oxygen permeation by analyzing dependence of the permeation flux
on membrane thickness.6,15,18 This quantity is, by definition, propor-
tional to jd, and should not depend on thickness when surface limi-
tations are negligible. In this situation, J(O2) is proportional to the
so-called ambipolar conductivity (␴amb), averaged for a given oxy-
gen partial pressure range
RT
16F2
RT
␴ ␴
RT
16F2
o
e
J O ͒ ϭ
␴
amb ϭ
ϭ
␴to 1 Ϫ t ͒
͑
o
͑
2
16F2 ␴o ϩ ␴
e
͓2͔
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
E127
Figure 2. Temperature dependence of the total conductivity of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ ceramics in air: ͑1͒ x ϭ 0; ͑2͒ x ϭ 0.2; ͑3͒
x ϭ 0.4. Curve
4
corresponds to the literature data10 on
La0.5Sr0.5Fe0.8Ga0.2O3Ϫ␦ compound, shown for comparison.
time, heating usually leads to similar conductivity values in the ma-
terials with different Sr content ͑e.g., see Ref. 22 and references
therein͒. Thus, the obtained conductivity values are in reasonable
agreement with the literature data.10
Thermal expansion coefficients ͑TECs͒ in air, calculated from
the dilatometric and high-temperature XRD data, are listed in Table
III. Within the limits of experimental error, thermal expansion of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ ceramics, measured by dilatometry, is very
similar to the expansion of crystal lattice. The TEC values ͑␣͒ vary
in a relatively narrow range (11.7-14.9) ϫ 106 KϪ1 at temperatures
below 800 K, and increase up to (19.5-26.4) ϫ 106 KϪ1 at 800-
1170 K. Such an increase, typical for numerous SrFeO3Ϫ␦-based
phases,14,15 is associated with oxygen losses on heating.20
Doping with gallium leads to decreasing thermal expansion of
La0.3Sr0.7Fe(Ga)O3Ϫ␦ , which may be explained either by the re-
duced losses of oxygen from gallium-containing specimens or by
smaller total concentration of iron cations. For instance, high TECs
of (La,Sr)CoO3Ϫ␦ are partially associated with numerous transitions
in the electronic state of cobalt.34 Respectively, decreasing concen-
tration of transition metal cations in the B sites should suppress the
lattice expansion when temperature increases.
Stability tests.—Testing of La0.3Sr0.7Fe(Ga)O3Ϫ␦ phases in
CO2-containing atmospheres demonstrated their satisfactory stabil-
Figure 1. SEM micrographs of the La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ ceramics: ͑a͒ x
ϭ 0; ͑b͒ x ϭ 0.2; ͑c͒ x ϭ 0.4.
Table II. Activation energy for p-type electronic conductivity of
La0.3Sr0.7Fe1ÀxGaxO3À␦ at 300-800 K in air.
Ea
ln(A0)
x
͑kJ/mol͒
͑S/cm͒
ing strontium content in the perovskites (La,Sr)MO3 , where M is a
transition metal such as Mn, Fe, or Co, usually results in the respec-
tive increase of the hole conductivity level and simultaneous de-
crease of the activation energy at low temperatures. At the same
0
0.20
0.40
14.3 Ϯ 0.1
13.7 Ϯ 0.1
12.61 Ϯ 0.05
9.6
Ϯ
Ϯ
Ϯ
0.5
16.6
27.3
0.4
0.2
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E128
Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
Table III. Average TECs, chemically induced expansion, and oxygen nonstoichiometry changes of La0.3Sr0.7Fe1ÀxGaxO3À␦ ceramics.
Average TECs in air
Dilatometry
High-temperature XRD
␣ ϫ 106
⌬L/Lb
1153 K
͑%͒
␣ ϫ 106
͑KϪ1
⌬␦a
1073 K
T
T
͑K͒
͑K͒
x
͒
͑KϪ1
͒
0
300-780
780-1140
300-920
920-1110
300-930
930-1100
300-770
770-1170
300-770
770-1170
300-770
770-1170
0.157
0.314
13.5
Ϯ
Ϯ
Ϯ
Ϯ
Ϯ
Ϯ
0.2
0.8
0.3
1
0.2
2
14.9 Ϯ 0.3
26.4 Ϯ 0.3
14.9 Ϯ 0.3
23.7 Ϯ 0.8
11.7 Ϯ 0.3
19.5 Ϯ 0.9
24.8
12.9
25
12.8
20
0.20
0.40
0.046
0.166
a ⌬␦ is the change in the oxygen nonstoichiometry under the oxygen partial pressure change from 21 kPa to approximately 3 Pa.
b ⌬L/L is the chemically induced expansion under the oxygen partial pressure change from 21 kPa to approximately 3 Pa, measured by dilatometric
technique.
ity with respect to interaction with carbon dioxide. As an example,
Fig. 3 presents TGA data on LSFG20 powder in purified air and in
a gas mixture containing CO2 . For comparison, the results of simi-
lar test of SrO are given. Before testing, the samples were annealed
in air at 1070-1250 K in the TGA cell, and then slowly cooled to 823
K. While formation of SrCO3 from strontium oxide occurs exten-
sively, no interaction between La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ and carbon
dioxide was detected by gravimetric analysis within the experimen-
tal error limits. The DTA tests for the La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ pow-
ders, prekept in the CO2 atmosphere during different periods of
time, did not reveal also any thermal effects on heating. This sug-
gests that adsorption of CO2 on the grain surface of the solid solu-
tions does not result in considerable formation of carbonates and in
gross decomposition of the perovskite phase.
On the other hand, IR absorption studies of the
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ powders indicate trace amounts of SrCO3
after keeping samples in the carbon dioxide atmosphere at room
temperature. The example is shown in Fig. 4, where a comparison is
given of the IR spectra for the La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦ powder be-
fore and after treatment in the pure CO2 atmosphere. The IR absorp-
tion band at ␯ Ӎ 860 cmϪ1 appearing in the spectrum of the
CO2-treated sample undoubtedly indicates a formation of minute
amounts of strontium carbonate.35 Therefore, one must conclude
that surface decomposition of the perovskite phase because of the
interaction with carbon dioxide, though slow, nonetheless takes
place at low temperatures. Similar behavior is characteristic of the
Sr(Fe,Ti)O3Ϫ␦ solid solutions.36 It is important, however, that com-
pact sintered ceramic samples of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ did not
show any evidence of the degradation even after several months of
storage in the CO2 atmosphere at room temperature.
Two other parameters determining the applicability of oxygen
membrane materials are related to an absence of time degradation of
the permeation fluxes and to isothermal chemically induced expan-
sion, caused by oxygen vacancy concentration changes under varia-
tions of the oxygen partial pressure. The degradation in oxygen per-
meability with time may originate from various reasons, including
reaction with gaseous species such as CO2 , phase transformations,
or ordering in the oxygen sublattice, resulting in a drop in the ionic
conductivity with time.8,11-14,16 Oxygen permeation tests of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ membranes did not show any significant
change in the ion transport at temperatures above 1070 K during
200-250 h ͑Fig. 5͒. This behavior is advantageous as compared to
that of Sr(Fe,Ti)O3Ϫ␦ perovskites or Ce(Gd)O2Ϫ␦ /La(Sr)MnO3Ϫ␦
composite membranes.14,16
Figure 3. Relative weight changes of La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦ and SrO in
atmospheric air purified from CO2 , and in a mixture of CO2 , O2 and N2 at
823 K. The ratio between CO2 , O2 , and N2 concentrations was 18:15:67
͑LSFG20͒, 19:17:64 ͑SrO͒.
Figure 4. IR absorption spectra of La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦ at room tempera-
ture: ͑1͒ after annealing in air at 1170 K for 2 h, and ͑2͒ after keeping in CO2
atmosphere at room temperature for 24 h. The spectra were obtained using
samples of the same powder before and after its treatment in CO2 .
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
E129
Figure 5. Time dependencies of the oxygen permeability of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ membranes under fixed oxygen partial pressure gra-
dients.
In agreement with the conclusion by Schwartz et al.,7 dilato-
metric tests showed that substituting iron with gallium in
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ leads to suppressing the chemically induced
expansion. Table I presents the values of the isothermal chemical
expansion (⌬L/L) of the ceramics with x ϭ 0 and 0.4, measured in
conditions when the oxygen partial pressure decreased from 21 kPa
͑atmospheric air͒ to about 3 Pa ͑pure argon͒. More detailed data on
the expansion resulting from the oxygen stoichiometry changes at
constant temperature will be published elsewhere. In this paper, the
enhanced dimensional stability of the Ga-substituted materials is
associated with smaller changes in the oxygen nonstoichiometry un-
der variations of oxygen partial pressure; an example is given in
Table I.
Oxygen permeability.—Figure 6a shows the dependence of the
oxygen permeation fluxes through La0.3Sr0.7FeO3Ϫ␦ membranes of
various thickness on the oxygen partial pressure gradient; the corre-
sponding values of the specific oxygen permeability are presented in
Fig. 6b. While the permeation fluxes decrease with increasing mem-
brane thickness, the oxygen permeability increases. This phenom-
enon suggests that the oxygen transport through La0.3Sr0.7FeO3Ϫ␦
ceramics is limited by both surface exchange and bulk ambipolar
conduction. Significant interphase exchange limitations of
oxygen permeation are typical for numerous SrFeO3Ϫ␦-and
SrCoO3Ϫ␦-based materials.14,15,18
Figure 6. Dependence of the ͑a͒ oxygen permeation fluxes and ͑b͒ specific
oxygen permeability of La0.3Sr0.7FeO3Ϫ␦ membranes on the oxygen partial
pressure gradient. Solid lines are for visual guidance only.
In contrast, similar results on La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦
(x ϭ 0.2-0.4) showed negligible effect of the oxygen surface ex-
change on oxygen permeability of Ga-containing materials ͑Fig. 7͒.
The J(O2) values of the La0.3Sr0.7Fe(Ga)O3Ϫ␦ ceramics are inde-
pendent of membrane thickness within the experimental error limits,
i.e., the integral form of the Wagner law is observed. Notice that the
scale in Fig. 7 is the same as that in Fig. 6b. Thus, the permeability
of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4) is predominantly limited
by the bulk ambipolar conductivity, which makes it possible to cal-
culate the ionic conductivity of La0.3Sr0.7Fe(Ga)O3Ϫ␦ from data on
oxygen permeation and total conductivity.
The difference in the behavior of undoped and Ga-doped
La0.3Sr0.7FeO3Ϫ␦ can be attributed to either increasing surface ex-
change rates or decreasing ionic conductivity when gallium substi-
tutes iron in the B sublattice. The obtained results do not allow
drawing a conclusive judgement on this matter, and both mentioned
reasons may be operable. Note also that the observed behavior of
La0.3Sr0.7(Fe,Ga)O3Ϫ␦ membranes seems to be in controversy with
the conclusion of Kim et al.,23 who suggested surface-limited oxy-
gen permeation through La0.5Sr0.5Fe0.8Ga0.2O3Ϫ␦ ceramics. Such a
contradiction may result, first, from the different strontium content
in the materials studied in this work and reported by Kim et al.23 It
is well known that both oxygen diffusivity and surface exchange
rates in (La,Sr)FeO3Ϫ␦ solid solutions are functions of La:Sr
ratio.37,38 Also, the conclusion23 was made comparing a simplified
model with experimental data for only one membrane thickness
without examining the dependence of permeation fluxes on the
thickness, and this may lead to uncertainties.
The
oxygen
fluxes
through
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦
(x ϭ 0-0.4) membranes are found to increase with x, whereas the
apparent activation energy for oxygen permeability decreases with
gallium additions ͑Fig. 8͒. For Ga-containing compositions
(x ϭ 0.2-0.4), the temperature dependence of the permeation is de-
termined by the activation energy for the bulk ambipolar conductiv-
ity ͑Eq. 2͒; the trends in the bulk ionic conductivity are discussed
below. In the case of undoped La0.3Sr0.7FeO3Ϫ␦ , however, both bulk
ion transport and surface exchange rate affect the temperature de-
pendence of the permeability. This can be confirmed comparing the
obtained apparent activation energy for oxygen permeability of
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E130
Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
Figure 8. Temperature dependence of the oxygen permeation flux density
through La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ membranes under fixed oxygen partial pres-
sure gradient. The membrane thickness is 1.00 Ϯ 0.01 mm.
ץj
␴
ϭ 4Fd
͓4͔
͓5͔
ͩ ͪ
ͱ1 Ϫ 4
amb
ץE
E→0
1
2
1
␴
amb
to ϭ
Ϫ
2
␴
where E is the Nernst voltage
E ϭ
RT
p2
ln
͓6͔
ͩ ͪ
4F
p1
Hereafter, the model constituted by Eq. 5-7 is referred to as model 1.
For the calculations by this model, only the permeation data corre-
sponding to the permeate-side oxygen pressures in the range from
approximately 5 to 21 kPa were used.
Figure 7. Dependence of the specific oxygen permeability of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ ceramics on the oxygen partial pressure gradient: ͑a͒
x ϭ 0.2; ͑b͒ x ϭ 0.4. Solid lines are for visual guidance only.
Table IV lists the oxygen ion transference numbers determined
from the FE results and calculated from the OP using Eq. 5-7.
Within the experimental error limits, both techniques give similar
values of the transference numbers. The values of to determined
from the FE data may be slightly underestimated, probably due to
non-negligible polarization resistance in the electrochemical cells
for FE measurements.36 The oxygen ion transference numbers of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4) in air at 1023-1223 K vary
from 4 ϫ 10Ϫ4 to 4 ϫ 10Ϫ2, increasing with gallium content and
temperature. The observed increase in to values with increasing x
results from the decrease of the p-type electron contribution to the
total conductivity when gallium content increases ͑Fig. 2͒. At the
same time, increasing temperature at constant gallium content leads
to an increase in the ion contribution while the hole contribution,
which exhibits metallic-like temperature behavior under fixed oxy-
gen pressure ͑Fig. 2͒, decreases. Consequently, when the oxygen
partial pressure is high and constant, the ion transference numbers
increase with temperature.
La0.3Sr0.7FeO3Ϫ␦ (161 Ϯ 4 kJ/mol) with the literature data on
surface exchange for (La,Sr)FeO3Ϫ␦ phases. For instance, the
activation energy for the surface exchange coefficient of
La0.75Sr0.25FeO3Ϫ␦ was calculated from the isotopic exchange data
to be equal to 185 Ϯ 27 kJ/mol, while the Ea value for the vacancy
diffusivity is as low as 114 Ϯ 23 kJ/mol.37 The activation energy
for the surface exchange of La0.6Sr0.4FeO3Ϫ␦ is reported equal to
131 Ϯ 4 kJ/mol.38 One can also notice that the Ea value for oxygen
permeation through La0.6Sr0.4FeO3Ϫ␦ membranes at 1123-1323 K is
found equal to 173 Ϯ 11 kJ/mol.38 Hence, the high activation en-
ergy for the oxygen permeability of La0.3Sr0.7FeO3Ϫ␦ ͑Fig. 8͒ may
partially result from the oxygen exchange limitations.
Ionic conductivity.—The fact that the permeation fluxes through
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4) are predominantly limited
by the bulk oxygen transport, made possible calculation of the ion
transference numbers and ionic conductivity from the data on oxy-
gen permeability and total conductivity ͑see Eq. 2͒
Figure 9 compares the values of oxygen ionic conductivity, de-
termined from the oxygen permeation results in oxidizing conditions
and calculated from the oxygen partial pressure dependencies of the
TC in reducing atmospheres. Notice that the estimations of the ionic
conductivity of La0.3Sr0.7FeO3Ϫ␦ in air, obtained from the OP data,
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
E131
Table
IV. Oxygen
ion
transference
numbers
of
La0.3Sr0.7Fe1ÀxGaxO3À␦ in air, calculated from oxygen perme-
ation and faradaic efficiency data.
to
T
͑K͒
x
OPa
FEb
8.5 ϫ 10Ϫ3
3.7 ϫ 10Ϫ3
1.7 ϫ 10Ϫ3
9.1 ϫ 10Ϫ4
4.1 ϫ 10Ϫ4
4.2 ϫ 10Ϫ2
2.8 ϫ 10Ϫ2
1.4 ϫ 10Ϫ2
7.5 ϫ 10Ϫ3
4.1 ϫ 10Ϫ3
0.20
1223
1173
1123
1073
1023
1223
1173
1123
1073
1023
-
-
1.4 ϫ 10Ϫ3
8.7 ϫ 10Ϫ4
-
3.7 ϫ 10Ϫ2
0.40
2.4 ϫ 10Ϫ2
-
-
-
a The transference numbers calculated from oxygen permeability and
total conductivity are averaged in the oxygen partial pressure range
from approximately 5 to 21 kPa.
b The tranference numbers were determined by the faradaic efficiency
methods under zero oxygen chemical potential in air.
are lower than the true values due to the significant surface ex-
change limitations to the permeation fluxes. In reducing environ-
ments, the total conductivity of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ can be ap-
proximated by the classical relation
o
n
o
Ϫ1/4
1/4
␴ ϭ ␴ ϩ ␴ p O ͒
ϩ ␴ p O ͒
͓7͔
͑
͑
o
2
2
p
Figure 10. Dependence of the total conductivity of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦
ceramics on the oxygen partial pressure at ͑a͒ 1223 and ͑b͒ 1023 K. The solid
lines correspond to the fitting results using Eq. 7 as regression model.
0
p
0
n
where ␴ and ␴ are the partial p-type and n-type electronic con-
ductivities extrapolated to the unit oxygen pressure, respectively,
and ␴ represents p(O2)-independent ionic contribution. Figure 10
o
demonstrates the adequacy of this model for the description of the
experimental data; a more detailed analysis of the transport proper-
ties of the title materials at low oxygen pressures can be found
elsewhere.24
When discussing the relationships between ionic transport in
La0.3Sr0.7Fe(Ga)O3Ϫ␦ in reducing and oxidizing conditions, one
should mention that decreasing oxygen chemical potential causes
various types of ordering in the oxygen sublattice of the title
materials.24 The oxygen depletion from La0.3Sr0.7FeO3Ϫ␦ at tem-
peratures below 1070-1120 K and oxygen partial pressures below
1-10 Pa results in formation of an orthorhombic brownmillerite-like
phase ͑space group Pmma͒, similar to the reduced form of the
Sr2LaFe3O9Ϫ␦ compound.39 Under the same conditions, substitution
of iron with gallium at x у 0.10 leads to stabilization of another
ordered perovskite-related phase having an orthorhombic structure
͑space group Pmmm͒.24 On heating, these reduced phases exhibit a
partial disordering, which is reflected by the decrease in the activa-
tion energy for ionic conductivity at approximately 1120 K.24
Neglecting the ordering in the oxygen sublattice, the ionic con-
ductivity of reduced La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4) is very
close to that of the oxidized phases ͑Fig. 9 and 11͒. This suggests
that the concentration of free oxygen vacancies, available for ionic
transport, is preserved sufficiently large in both disordered and par-
tially ordered phases. As a result, the ion conductivity is rendered
Figure 9. Comparison of the oxygen ionic conductivity of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ , determined from the data on OP at the oxygen
partial pressures of 5-21 kPa using Eq. 5-7, and calculated from the oxygen
pressure dependencies of the TC using Eq. 7, as demonstrated by Fig. 10.
The apparent ionic conductivity of LSF ceramics was calculated from the
permeation data on the membrane with d ϭ 1.00 mm and is significantly
underestimated due to the surface exchange limitations.
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E132
Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
The values of ionic conductivity of La0.3Sr0.7FeO3Ϫ␦ in air, as
estimated from the oxygen permeation data using Eq. 5-7, are close
to those of La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦ at 1173-1223 K ͑Fig. 9 and 11͒.
However, taking into account the surface limitations of oxygen
transport through La0.3Sr0.7FeO3Ϫ␦ membranes, one can conclude
that the oxygen ionic conductivity of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦
Ͻ
(x
0.3) at temperatures above 1170 K decreases with gallium
additions in both oxidizing and reducing conditions. Exact reasons
for such phenomenon are still unclear. One could mention that simi-
lar behavior is characteristic of La(Co,Ga)O3Ϫ␦ solid solutions,
where the ionic transport in the intermediate compositions is con-
siderably lower than in both parent compounds, LaCoO3Ϫ␦ and
42
LaGaO3Ϫ␦
.
A possible explanation may be related to significant
difference in the covalence of Ga-O and Fe-O chemical bonds,
which results in strong local distortions in the crystal lattice around
the guest gallium cations. Such distortions may favor trapping of the
oxygen anions in the neighborhood of gallium cations, thus decreas-
ing concentration of the mobile oxygen ions.
Finally, some comments should be made on the large difference
in ionic conductivity of La0.3Sr0.7FeO3Ϫ␦ in reducing and oxidizing
atmospheres ͑Fig. 9͒. The main and obvious reason is that the ap-
parent ionic conductivity of undoped lanthanum-strontium ferrite at
high oxygen pressures is underestimated because of the surface ex-
change effect on the oxygen permeation. However, the observed
large difference in ␴o values ͑Fig. 9͒ may also indicate that the ionic
conductivity of the reduced form of La0.3Sr0.7FeO3Ϫ␦ is greater than
that of the oxidized form, neglecting the ordering of oxygen vacan-
cies at low oxygen chemical potentials. This difference may result
from the larger variation of the oxygen content under decreasing
oxygen pressure ͑␦⌬͒ in the parent lanthanum-strontium ferrite than
in the gallium-doped derivatives. Indeed, the oxygen content change
at reduction of La0.3Sr0.7FeO3Ϫ␦ , as measured by coulometric titra-
tion, is significantly larger than that in the Ga-doped derivatives
͑e.g., Table I and Ref. 24͒. Respectively, and in contrast to Ga-
containing compositions, this favors the ion conductivity in
La0.3Sr0.7FeO3Ϫ␦ ceramics to increase with the decrease in partial
pressure of oxygen.
Figure 11. Composition dependence of the oxygen ionic conductivity of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ , calculated from the oxygen permeation data in oxi-
dizing atmosphere and from the total conductivity at reduced oxygen pres-
sures. Solid lines are for visual guidance only.
essentially independent on oxygen content and ordering phenomena
in the oxygen partial pressure range from 10Ϫ18 to 21 kPa. As com-
pared to the composition with x ϭ 0.2, the La0.3Sr0.7Fe0.6Ga0.4O3Ϫ␦
ceramics exhibit a slightly higher ionic conductivity and a lower
activation energy for oxygen transport ͑Fig. 11 and Table V͒. Most
probably, this difference may be attributed to the decrease in the
nominal ratio La:Sr ϭ 3/7 in the perovskite, caused by segregation
of SrLaGa3O7 phase at x ϭ 0.4. For instance, an increase in ionic
conduction and a decrease in the activation energy is observed in the
La1ϪySryCoO3Ϫ␦ (x ϭ 0.4-0.8) system when the La:Sr concentra-
tion ratio decreases.40,41
Oxygen pressure dependence of the permeation fluxes.—Since
the ionic conductivity of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4) is
essentially independent of oxygen pressure, the p(O2)-dependencies
of the oxygen permeability and ambipolar conductivity of these ma-
terials are determined by the electronic conduction. Therewith, the
ion transference numbers in air are relatively small (to р 4
ϫ 10 Ϫ 2) and decrease when x or temperature decreases, Table II.
In these conditions, one could assume that the permeation fluxes are
limited only by oxygen ionic conductivity. Such an assumption is
quite common in the literature dealing with oxygen permeability of
Fe- or Co-containing perovskite-type mixed conductors.3,43,44 How-
ever, if this assumption was valid for the title materials, their spe-
cific oxygen permeability ͑Eq. 2͒ should be independent of oxygen
pressure, which contradicts the experimental data ͑Fig. 7͒. There-
fore, the electronic conductivity definitely affects oxygen perme-
ation and ambipolar conductivity of La0.3Sr0.7Fe(Ga)O3Ϫ␦
membranes.
Analysis of the oxygen partial pressure dependencies of the total
conductivity shows that, within the p(O2) range where oxygen per-
meation was studied ͑0.9-21 kPa͒, the total conductivity is approxi-
mately proportional to p(O2)1/4, in agreement with the classical
model for p-type electronic conduction. Figure 12 presents some
examples of the experimental data on total conductivity in oxidizing
conditions. The fitting results given in Fig. 12 were obtained using
the power model
Table V. Activation energy for oxygen ionic conductivity of
La0.3Sr0.7Fe1ÀxGaxO3À␦ at 1023-1223 K.
Ea
͑kJ/mol͒
Calculated from
Calculated from the total
conductivity datab
x
oxygen permeation dataa
117 Ϯ 5c
-
134 Ϯ 5
119 Ϯ 4
79 Ϯ 3
100 Ϯ 7
125 Ϯ 10
121 Ϯ 4
0
0.10
0.20
0.40
a The values of the ionic conductivity calculated from the oxygen per-
meation data refer to the oxygen partial pressure range from 5 to 21
kPa.
b The values of the ionic conductivity, calculated from the oxygen
pressure dependence of the total conductivity, correspond to the
oxygen pressure range from 10Ϫ12 to 10Ϫ14 Pa.
␴ ϭ k p O ͒1/4 ϩ k2
͓8͔
͑
1
2
c The value of the activation energy for ionic conductivity of
La0.3Sr0.7FeO3Ϫ␦ ͑1173-1223 K͒, estimated from the permeation
data, is affected by the activation energy for oxygen surface ex-
change.
where k1 and k2 are the regression parameters. The simple
␴ ϳ p(O2)1/4 dependence, which follows from the experiment,
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
E133
Figure 13. Dependence of the oxygen permeation fluxes of
La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦ membranes on the oxygen partial pressure gradient.
The solid lines correspond to the fitting results using Eq. 14 as regression
model.
Finally, in the theoretical case when the ionic and electronic con-
ductivities both are independent on oxygen pressure ͑i.e., k1 → 0,
permeation flux29
␴ ϭ k2͒, Eq. 12 transforms in the classical formula for oxygen
e
Figure 12. Dependence of the total conductivity of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦
ceramics on the oxygen partial pressure at ͑a͒ 1123 and ͑b͒ 1223 K. The solid
lines correspond to the fitting results using Eq. 8 as regression model.
RT
␴ ␴
p2
p1
o
e
j ϭ
ln
͓13͔
16F2d ␴o ϩ ␴
e
makes it possible to find an analytical solution of the Wagner equa-
tion for the oxygen permeation flux, valid for La0.3Sr0.7Fe(Ga)O3Ϫ␦
membranes in oxidizing conditions.
The permeation process limited by the bulk ambipolar conduc-
tion can be described as
For La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4) membranes in oxi-
dizing conditions, the constant, m, is approximately equal to 1/4
͑Fig. 12͒. Respectively, the results of fitting La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦
permeation data ͑Fig. 13͒ were obtained using the model derived
from Eq. 13
p
2
RT
␴ ␴
o e
k1p1/4 ϩ ␴
j ϭ
ץ ln p O ͒
͔
2
͓9͔
͓
͑
RT␴
4F2d
͵
16F2d
␴o ϩ ␴
o
o
o
2
e
p
j ϭ
ln
ϩ k3
͓14͔
1
k1p1/4 ϩ ␴
1
When the ionic conductivity is independent of oxygen chemical
potential, and the electronic conductivity follows to the power de-
pendence on the oxygen partial pressure, similar to Eq. 9
where the regression parameters included ␴o , k1 , and k3 . For all
permeation data on La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4) at 1023-
1223 K, the correlation coefficient of the model Eq. 2 ͑referred to as
model 2͒ was no less than 0.993. The errors in the regression pa-
rameters of the model 2, excepting k3 , did not exceed 8%, which is
comparable with the experimental error. As a rule, the parameter k3
was statistically insignificant, but its use was necessary to provide a
better quality of fitting. Figure 14 presents a comparison between
ionic conductivity values, calculated from the permeation fluxes us-
ing models 1 and 2 and derived from the total conductivity data. In
general, the adequacy of model 2 is quite good, supporting the
above conclusion that the variations of the oxygen partial pressure
m
␴ ϭ k p O ͒ ϩ k2
͓10͔
͑
e
1
2
with m being the constant, the analytical solution of Eq. 10 is as
follows
k1pm2 ϩ ␴o ϩ k2
k1pm1 ϩ ␴o ϩ k2
RT
␴
p2
p1
o
j ϭ
␴
ln
ϩ k2m ln
ͫ
ͬ
o
16F2dm ␴o ϩ k2
͓11͔
has no effect on ionic conductivity of La0.3Sr0.7Fe(Ga)O3Ϫ␦
.
This equation may be simplified if k2 → 0
When discussing these results, it should be emphasized that the
ratio between ionic and electronic contributions to the total conduc-
tivity of La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦ is relatively small: the ion transfer-
ence numbers of this material in the studied temperature range are in
k1pm ϩ ␴
RT␴
o
o
o
2
j ϭ
ln
͓12͔
16F2dm
k1pm ϩ ␴
1
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E134
Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
Oxygen permeability of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ ceramics, mea-
sured in the range of oxygen partial pressures at the membrane
permeate side from 0.9 to 21 kPa, increases with increasing x at
1023-1223 K. While the permeability of La0.3Sr0.7FeO3Ϫ␦ is limited
by both bulk ionic conduction and surface exchange rates, the per-
meation fluxes through Ga-containing membranes are determined
predominantly by the bulk ambipolar conductivity. The ion transfer-
ence numbers of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0.2-0.4) at oxygen
partial pressures, close to atmospheric air, vary from 4 ϫ 10Ϫ4 to
4 ϫ 10Ϫ2, increasing with gallium content and temperature. Al-
though there is a great difference between ionic and p-type elec-
tronic conductivities in oxidizing atmospheres, the electronic con-
duction was shown to affect oxygen permeation fluxes.
Incorporation of small amounts of Ga (x р 0.2) into the crystal
lattice of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ results in decreasing oxygen ionic
conductivity at 1173-1223 K. Further doping leads to a greater ionic
conduction and lower activation energy for ionic transport due to the
decrease in La:Sr concentration ratio, caused by segregation of
SrLaGa3O7 phase. Ionic conductivity of La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦
(x ϭ 0.2-0.4) is essentially independent of oxygen chemical poten-
tial, oxygen nonstoichiometry, and structural transformations in the
oxygen partial pressure range from 10Ϫ14 to 0.21 kPa.
Acknowledgments
This research was partially supported by FCT ͑Praxis, Portugal͒,
the University of Aveiro, the Belarus Ministry of Education and
Science, and the Russian Foundation for Basic Research.
Figure 14. Comparison of the oxygen ionic conductivity of
La0.3Sr0.7Fe0.8Ga0.2O3Ϫ␦ , calculated by the two different models from the
oxygen permeation data, and derived from the oxygen partial pressure de-
pendencies of the total conductivity ͑see text͒.
The University of Aviero assisted in meeting the publication costs of this
article.
References
1. H. J. M. Bouwmeester and A. J. Burgraaf, in Fundamentals of Inorganic Mem-
brane Science and Technology, A. J. Burggraaf and L. Cot, Editors, p. 435,
Elsevier, Amsterdam ͑1996͒.
2. T. J. Mazanec, in Ceramic Membranes I, H. U. Anderson, A. C. Khandkar, and M.
Liu, Editors, PV 95-24, p. 16, The Electrochemical Society Proceedings Series,
Pennington, NJ ͑1997͒.
3. Y. Teraoka, H. M. Zhang, K. Okamoto, and N. Yamazoe, Mater. Res. Bull., 23, 51
͑1988͒.
4. R. M. Thorogood, R. Srinivasan, T. F. Yu, and M. P. Drake, U.S. Pat. 5,240,480
͑1993͒.
5. G. Dixon, W. R. Moser, and Y. H. Ma, Ind. Eng. Chem. Res., 33, 3015 ͑1994͒.
6. V. V. Kharton, A. A. Yaremchenko, A. V. Kovalevsky, A. P. Viskup, E. N. Naumov-
ich, and P. F. Kerko, J. Membr. Sci., 163, 307 ͑1999͒.
7. M. Schwartz, J. White, and A. Sammels, Int. Pat. Appl. PCT WO 97/41060 ͑1997͒.
8. S. Pei, M. S. Kleefisch, T. P. Kobylinsky, J. Faber, C. A. Udovich, V. Zhang-
McCoy, B. D. Dabrowski, U. Balachandran, R. L. Mieville, and R. B. Poeppel,
Catal. Lett., 30, 201 ͑1995͒.
the range from 4 ϫ 10Ϫ4 to 9 ϫ 10Ϫ3 ͑Table II͒. However, the
p-type electronic conductivity still affects the oxygen permeation
fluxes. A particular result of this influence refers to the negative
slope of the dependencies of the oxygen permeability, J(O2), on
oxygen partial pressure gradient ͑Fig. 7͒. Therefore, one can con-
clude that it may be incorrect to ascribe the oxygen pressure depen-
dence of the transport coefficients ͑proportionality coefficients be-
tween permeation flux and chemical potential gradient͒ in the
Wagner-type equations only to p(O2)-dependence of the permeation
rate-limiting parameters, such as ionic conductivity or surface
exchange rate.23,44 The oxygen pressure dependencies of the trans-
port properties, derived from the permeation data, may be affected
by the electronic conduction, even in case of materials with pre-
dominant electronic conductivity.
9. T. J. Mazanec, T. L. Cable, J. G. Frye, and W. R. Kliewer, U.S. Pat. 5,591,315
͑1997͒.
10. Q. Ming, M. D. Nersesyan, A. Wagner, J. Ritchie, J. T. Richardson, D. Luss, A. J.
Jacobson, and Y. L. Yang, Solid State Ionics, 122, 113 ͑1999͒.
11. H. Kruidhof, H. J. M. Bouwmeester, R. H. E. van Doorn, and A. J. Burggraaf, Solid
State Ionics, 63-65, 816 ͑1993͒.
12. M. F. Carolan, P. N. Dyer, S. M. Fine, J. M. LaBar, and R. M. Thorogood, U.S. Pat.
5,269,822 ͑1993͒.
13. M. V. Patrakeev, I. A. Leonidov, E. B. Mitberg, A. A. Lakhtin, V. G. Vasiliev, V. L.
Kozhevnikov, and K. R. Poeppelmeier, Ionics, 5, 444 ͑1999͒.
14. V. V. Kharton, F. M. Figueiredo, A. V. Kovalevsky, A. P. Viskup, E. N. Naumovich,
J. R. Jurado, and J. R. Frade, Defect Diffus. Forum, 186-187, 119 ͑2000͒.
15. V. V. Kharton, V. N. Tikhonovich, Li. Shuangbao, E. N. Naumovich, A. V. Kova-
levsky, A. P. Viskup, I. A. Bashmakov, and A. A. Yaremchenko, J. Electrochem.
Soc., 145, 1363 ͑1998͒.
16. V. V. Kharton, A. V. Kovalevsky, A. P. Viskup, F. M. Figueiredo, A. A. Yarem-
chenko, E. N. Naumovich, and F. M. B. Marques, J. Electrochem. Soc., 147, 2814
͑2000͒.
17. V. V. Kharton, E. N. Naumovich, and A. V. Nikolaev, J. Membr. Sci., 111, 149
͑1996͒.
18. A. V. Kovalevsky, V. V. Kharton, V. N. Tikhonovich, E. N. Naumovich, A. A.
Tonoyan, O. P. Reut, and L. S. Boginsky, Mater. Sci. Eng., B, 52, 105 ͑1998͒.
19. V. V. Kharton, A. P. Viskup, E. N. Naumovich, and F. M. B. Marques, J. Mater.
Chem., 9, 2623 ͑1999͒.
20. M. V. Patrakeev, I. A. Leonidov, A. A. Lakhtin, E. B. Mitberg, and V. L. Kozhevni-
kov, Zh. Fiz. Khim., 70, 611 ͑1996͒, in Russian.
21. V. L. Kozhevnikov, I. A. Leonidov, M. V. Patrakeev, E. B. Mitberg, and K. R.
Poeppelmeier, J. Solid State Chem., 158, 320 ͑2001͒.
Conclusions
Dense
ceramic
membranes
of
perovskite-type
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ (x ϭ 0-0.40) were prepared by the stan-
dard ceramic synthesis technique. The solid solution formation
range in this system corresponds to the x values from 0 to 0.3;
further increase in the gallium concentration leads to segregation of
SrLaGa3O7 . The total conductivity of La0.3Sr0.7Fe(Ga)O3Ϫ␦ , pre-
dominantly p-type electronic in oxidizing conditions, decreases
when insulating Ga cations are incorporated into the lattice of
lanthanum-strontium ferrite. Thermal expansion coefficients change
within (11.7-14.9) ϫ 106 and (19.5-26.4) ϫ 106 KϪ1 below 800 K
and between 800-1170 K, respectively, and decrease with the in-
crease in gallium content. TGA/DTA and oxygen permeation tests
demonstrated a satisfactory stability of the title materials in various
environments. The introduction of gallium in the crystal lattice
is found to suppress the isothermal volume changes of
La0.3Sr0.7Fe1ϪxGaxO3Ϫ␦ at variations of oxygen partial pressure that
is related to the narrowing of the oxygen homogeneous range with
gallium doping.
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 149 ͑4͒ E125-E135 ͑2002͒
E135
22. V. V. Kharton, A. A. Yaremchenko, and E. N. Naumovich, J. Solid State Electro-
chem., 3, 303 ͑1999͒.
34. K. Huang, H. Y. Lee, and J. B. Goodenough, J. Electrochem. Soc., 145, 3220
͑1998͒.
23. S. Kim, S. Wang, X. Chen, Y. L. Yang, N. Wu, A. Ignatiev, A. J. Jacobson, and B.
Abeles, J. Electrochem. Soc., 147, 2398 ͑2000͒.
35. IR Spectra of Inorganic Compounds, Vol. 4, R. A. Nyquist and R. O. Kagel,
Editors, Academic Press, San Diego ͑1997͒.
24. I. A. Leonidov, V. L. Kozhevnikov, E. B. Mitberg, M. V. Patrakeev, V. V. Kharton,
and F. M. B. Marques, J. Mater. Chem., 11, 1202 ͑2001͒.
25. M. V. Patrakeev, E. B. Mitberg, A. A. Lakhtin, I. A. Leonidov, V. L. Kozhevnikov,
and K. R. Poeppelmeier, Ionics, 4, 191 ͑1998͒.
36. V. V. Kharton, A. P. Viskup, A. V. Kovalevsky, F. M. Figueiredo, J. R. Jurado, A. A.
Yaremchenko, E. N. Naumovich, and J. R. Frade, J. Mater. Chem., 10, 1161 ͑2000͒.
37. T. Ishigaki, S. Yamauchi, K. Kishio, J. Mizusaki, and K. Fueki, J. Solid State
Chem., 73, 179 ͑1988͒.
26. E. B. Mitberg, M. V. Patrakeev, A. A. Lakhtin, I. A. Leonidov, V. L. Kozhevnikov,
and K. R. Poeppelmeier, J. Alloys Compd., 274, 103 ͑1998͒.
38. J. E. ten Elshof, Ph.D. Thesis, University of Twente, Enschede, The Netherlands
͑1997͒.
27. A. A. Yaremchenko, V. V. Kharton, A. P. Viskup, E. N. Naumovich, N. M. Lap-
chuk, and V. N. Tikhonovich, J. Solid State Chem., 142, 325 ͑1999͒.
28. V. V. Kharton, A. A. Yaremchenko, A. P. Viskup, G. C. Mather, E. N. Naumovich,
and F. M. B. Marques, Solid State Ionics, 128, 79 ͑2000͒.
29. H.-H. Moebius, Extended Abstracts of the 37th Meeting of the International Soci-
ety of Electrochemistry, Vol. 1, p. 136, Aug 25-30, 1986.
30. JCPDS Powder Diffraction File, card 45-0637, International Center Diffraction
Data.
31. P. D. Battle, T. C. Gibb, and P. Lightfoot, J. Solid State Chem., 84, 271 ͑1990͒.
32. N. Trofimenko and H. Ullmann, Solid State Ionics, 118, 215 ͑1999͒.
33. N. Ramadass, Mater. Sci. Eng., B, 36, 231 ͑1978͒.
39. P. D. Battle, T. C. Gibb, and P. Lightfoot, J. Solid State Chem., 84, 237 ͑1990͒.
40. R. H. E. van Doorn, Ph.D. Thesis, University of Twente, Enschede, The Nether-
lands ͑1996͒.
41. V. V. Kharton, E. N. Naumovich, A. A. Vecher, and A. V. Nikolaev, J. Solid State
Chem., 120, 128 ͑1995͒.
42. V. V. Kharton, A. P. Viskup, E. N. Naumovich, and N. M. Lapchuk, Solid State
Ionics, 104, 67 ͑1997͒.
43. J. A. Lane, S. J. Benson, D. Waller, and J. A. Kilner, Solid State Ionics, 121, 201
͑1999͒.
44. C. H. Chen, H. Kruidhof, H. J. M. Bouwmeester, and A. J. Burggraaf, J. Appl.
Electrochem., 27, 71 ͑1997͒.
Downloaded on 2015-06-18 to IP 142.103.160.110 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).